2010 Parallels. All Rights Reserved. 2

Transcription

1 Delivering Extraordinary Density for Cloud Service Providers Parallels Shared Hosting NG Service Module and Intel Xeon 5500 Processor Series-Based Servers Abstract: The goal of this paper is to provide detail on how to best deploy highly efficient, next generation Shared Web Services to enable Cloud Service Providers to deploy a more scalable and resilient Intel infrastructure, while at the same time delivering higher service levels at lower costs.

2 Table of Contents 1. KEY FACTORS IN SHARED WEB HOSTING DEPLOYMENTS SHARED WEB HOSTING IN THE PUBLIC CLOUD MARKET SHARED WEB HOSTING NG ARCHITECTURE SHARED WEB HOSTING INTEL TEST BED CHARACTERIZING SHARED WEB HOSTING WORKLOADS LINUX SHARED HOSTING NG WORKLOAD RESULTS BENEFIT TO CLOUD SERVICE PROVIDERS Parallels. All Rights Reserved. 2

3 1. KEY FACTORS IN SHARED WEB HOSTING DEPLOYMENTS The Need for the Next Generation of Web Hosting Shared Hosting has become a commodity. Many cloud service providers offer it as an entry level service to up-sell other offerings. Shared Web Hosting has low differentiation between providers and the margins associated with it are also low. Meanwhile, end users expect these commodity services to have the same high availability as their service. The more added value services customers consume, the higher their expectations are that Shared Web Hosting should be highly available and reliable, even if they pay virtually nothing for the service. In an effort to grow their customer base, large cloud service providers, like Google or Microsoft or other Mass Market Hosters, use their economies of scale to offer entry level Shared Web Hosting at very low prices. They can then offer additional feature rich services like business class and unified communications, as such value added services bring much richer profit margins. The goal of this paper is to enable cloud service providers at any level to be able to deploy a more scalable and resilient Shared Web Hosting infrastructure, while at the same time delivering higher service levels at lower costs. The Challenge of Cloud Service Providers In order to meet the high expectations of Shared Web Hosting end users, cloud service providers need to be extremely efficient in how they offer Shared Web Hosting in order to make it a profitable service. Based on Parallels experience, there are five things a cloud services provider can do to increase the efficiency of Shared Web Hosting: Lower costs by increasing the density of sites per server Increase average revenue per user by offering more reliable and available Shared Hosting service and/ or up-sell the customer to purchase additional services such as Hosted Exchange Improve service level management and segregate customers by usage or service plans across the server farm Increase predictability of capacity planning for the service and understand the relationship of capacity to service level performance Automate the user experience and empower the end user Lower Costs by Increasing the Density of Websites per Server Increasing the density of sites per shared hosting server can be difficult to accomplish. Application limitations, lack of isolation, CPU performance, memory capacity, server storage, and server I/O capacity are common limitations. Shared Web Hosting server farms must be carefully optimized to maximize the number of site per server while, at the same time, achieving the required performance levels. Lack of redundancy, however, increases the potential for service outages as server or networking failures occur more frequently in low-end servers that do not offer high reliability components. If density plus redundancy can be achieved, datacenter cost savings can be realized. Increase Average Revenue per User by Offering Other Services A common sales strategy for shared hosting is to get a large base of customers and then sell additional services to increase the average revenue per user. Common additional services include domain names, web site builder applications, SSL certificates, business class services, and Software as a Service. Shared Hosting plans are generally separated by commodity resources like storage and Internet bandwidth Parallels. All Rights Reserved. 3

4 Improve Service Level Management Shared Web Hosting service levels are generally based on a best effort basis. Cloud service providers do their best to have high system availability, but they do not offer guaranteed Service Level Agreements and/or the means to measure service availability to compensate end users should an outage occur. Since there is no SLA, heavy service users can consume more than their fair share of resources, which impacts other users and increases the potential of service churn due to a poor experience. Increase Predictability of Capacity and Availability Planning Managing the Shared Web Hosting service availability and capacity is also very important in providing a positive customer experience. Service growth over time is generally very predictable and cloud service providers who monitor their capacity usage can predict when additional resources will be required such as additional servers, storage capacity or network switches. Cloud service providers who do not actively monitor service resource consumption may realize too late that additional resources are needed, and this can create poor performance for weeks until additional resources can be added. Automate the User Experience and Empower the End User Automating the user experience is extremely important as end users expect cloud services to be ondemand. End users want to be in control of their account and expect to have the ability to control service features directly via a control panel and not have to open a trouble ticket or place a call to the cloud services provider. The margins are very thin with Shared Web Hosting services and a single support call can eliminate the margins associated with that account for an entire year. Parallels Operations Automation Linux Shared Hosting NG Module addresses these five key areas, enabling cloud service providers who supply Shared Web Services to operate a higher density, more efficient and available service. While only in beta at the present time, this white paper will explore best practices for deploying the Linux Shared Hosting NG Service Module solution on Intel Xeon 5500 processor series-based servers, including an architectural overview, test-bed hardware requirements, workload characterization, and the test-bed performance results obtained. 2. SHARED WEB HOSTING IN THE PUBLIC CLOUD MARKET Shared Web Hosting Background Shared web hosting service or virtual hosting service refers to a web hosting service where many websites reside on one web server connected to the Internet. Each site sits on its own partition, or section on the server to keep it separate from other sites. This is generally the most economical option for hosting, as many people share the overall cost of server maintenance. The hosting service must include system administration since it is shared by many users; this is a benefit for users who do not want to deal with admin functions, but a hindrance to power users who want more control. In general shared hosting will be inappropriate for users who require extensive software development outside what the hosting provider supports. Almost all applications intended to be on a standard web server work fine with a shared web hosting service. Shared hosting usually has usage limits and most hosting providers do not have extensive reliability features in place. Shared hosting typically offers a web-based control panel system, such as the Parallels Plesk Panel product suite, as an added value service. In shared hosting, the provider is generally responsible for managing servers, installing server software, security updates, technical support, and other aspects of the service. It is typically based on the LAMP stack. This is driven by the low cost of open source software, though some providers offer Microsoft Windows-based solutions. The initial release of Parallels Linux Shared Hosting NG Service Module is focused on LAMP based websites Parallels. All Rights Reserved. 4

5 The current approach for Parallels Shared Web Hosting is depicted in the following diagram, where the website data is stored directly on web servers and there are no load balancers to control the assignment websites to web servers. If a web server fails, all websites assigned to that server must be restarted. With 1000 to 2000 websites per server, this can take ~20 to ~30 minutes. 3. SHARED WEB HOSTING NG ARCHITECTURE Linux Shared Hosting NG Service Module Main Components The main components consist of Load Balancers, Linux Web Servers (with Apache), Linux Database Servers (with MySQL or PostgreSQL), and a NFS Share Storage Server. The diagram below details the architectural flow. Load Balancers Required for High Availability, Dynamic Load Redistribution and Seamless Capacity Management Any load balancing devices that provide direct routing and sticky sessions are supported as well Most service providers are expected to deploy hardware load balancers Linux LVS is used in the Linux Shared Hosting NG Service Module test-bed as it is an easy to configure, cost-effective solution for testing purposes 2010 Parallels. All Rights Reserved. 5

6 Web Servers Every web-server can process requests for any site Apache was chosen to utilize its specific features (.htaccess ) widely used in existing websites and required for seamless migration Web-Servers also handle FTP and SSH connections Database Servers Customers MySQL or PostgreSQL databases are handled by separate servers NFS Shared Storage Server NFS3 file systems are used to store sites content All file systems are mounted to all web servers Configuration Database Configuration of all sites handled by web cluster system are stored in a Central Configuration Database on the NFS Shared Storage Server SHNG Cluster Components HTTP(s) Requests Internet Load Balancer Apache Pro FTPD sshd Apache Pro FTPD sshd SHNG Module SHNG NSS plugin... SHNG Module SHNG NSS plugin Customersí Darabases MySQL PgSQL SHNG Caching Service Web-Server 1 SHNG Caching Service Web-Server N DB Servers Sitesí content (NFS3) File system 1 File system m Conf. DB Replica... Sitesí configuration (SQL) Conf. DB NFS Shared Storage Configuration Database 2010 Parallels. All Rights Reserved. 6

7 Linux Shared Hosting NG Implementation Details Parallels and Intel engineers built a test-bed based on Intel Xeon 5500 Processor Series-based servers. The key implementation details for the Linux Shared Hosting NG Service Module on the Intel test-bed include a set of functions that are diagramed by the web requests processing flow. Load Balancer and Shared IPs Sites are grouped by IP, about 1000 sites per IP Direct routing is used any web server can answer requests for any IP Sticky sessions are used for effective caching Website Isolation Dynamic content requests are processed through on demand standalone cgig processes that fork and create for each site an Parallels Containers-based Light Weight Environment (LVE) LVE context is limited by CPU, RAM, user UID and GID All dynamic requests to one site come to the same LVE and share CPU and RAM limits LVE is removed by kernel once all processes for a given site are finished Static content requests are processed by Apache Request rates and bandwidth are throttled in Apache All limits are stored per site in central configuration database Configuration Database and Caching Service The Linux Shared Hosting NG Service Module Caching Service is used to retrieve sites configuration data; this service caches data in RAM, and cached entries are removed if they are changed in Configuration Database The Linux Shared Hosting NG Service Module Caching Service automatically switches to a replica server if the main Configuration Database is inaccessible NFS3 Shared Storage and Content Caching HTTP(s) Requests Processing Diagram Standard Linux kernel NFS file system cache is used to improve performance Web Server Log Writing Through the Linux Shared Hosting NG Module Web Server Site Logs are stored in a site content directory on the Shared Storage Server Site Log writing is buffered in the Linux Shared Hosting NG Service Module; there is one common buffer for all sites handled by one server, which is flushed to disk once every 10 seconds or if the size reaches 16MB Apache SHNG Module SHNG Caching Service Sitesí content (NFS3) Dynamic Requests cgid 1..c [root] LVE (site-id) [uid, gid] Web-Server 1 IP 1...IP n File system1 Internet... HTTP(s) Requests IP 1...IP n Load Balancer... File systemm Apache SHNG Module SHNG Caching Service Conf. DB Replica Dynamic Requests cgid 1..c [root] LVE (site-id) [uid, gid] Web-Server N IP 1...IP n Conf. DB NFS Shared Storage Configuration Database 2010 Parallels. All Rights Reserved. 7

8 Shared Hosting NG Cloud Linux Architecture To achieve isolation and granular control of individual websites, the Parallels Linux Shared Hosting NG Module will be implemented on a new Linux offering based on CentOS. Called Cloud Linux, this new variant will offer lightweight virtual environments (LVEs) to control CPU and RAM allocations on a per website basis. The LVE technology is based on Parallels Virtuozzo Containers. The key differences between a standard Linux OS and Cloud Linux are detailed in the diagram below. Standard Linux OS (e.g. CentOS) Standard OS Multiple web sites per server Any web site can drain resources for a variety of reasons Single website can overload the server Increases risk and lowers efficiency of server Cloud Linux Resource isolation via lightweight virtual environment (LVE) Prevents individual web sites from using too much resources Single website cannot overload the server Lowers risk, increases efficiency and improves server performance Web sites Web sites Standard Linux Host OS Cloud Linux Hardware Hardware In a standard Linux OS kernel there is no ability to provide for resource control. This shared resource pool approach leads to websites which can over-consume precious compute and memory resources, impacting the performance of every website on a given server. With Parallels Lightweight Virtual Environments, websites on Cloud Linux will be able to apply CPU and memory limits on a per site basis (as noted by the dotted red lines in the diagram above). The Parallels Linux Shared Hosting NG Service Module will enable dynamic adjustment of service levels, allowing a customer to change their service plan to guarantee more resources if they find their website is sluggish and requires a higher level of performance Parallels. All Rights Reserved. 8

9 4. SHARED WEB HOSTING INTEL TEST BED The Shared Web Hosting test bed is comprised of Load Balancers, Web Servers, Shared Storage, and Parallels Operations Automation management server. The diagram below details the overall logical layout. The physical layout of the Linux Shared Hosting NG Service Module test bed consists of 16 Nehalembased EP Servers (Xeon X5570 dual socket/16gb RAM/500GB Local Disk) with CentOS 5.4 x86_64; Niantic 10GbE NICs were used for internal connections Parallels. All Rights Reserved. 9

10 NFS storage was configured using an Intel Timber Creek Dual CPU Storage Server with 8 GB memory, 12 SATA 1TB disk, RAID controller and Open-E NAS Storage software. Each server was also connected to 1GbE switch for management interface (not mentioned on diagram). A second test bed was created to compare the performance improvements found in Intel Xeon 55xx servers over Intel Xeon 54xx based servers (shown as al01-al03 on the diagram above) configured with 2xIntel Xeon E5450 3GHz, 8GB of RAM, 250GB SATA HDD. This second cluster shares same MySQL servers, NFS storage, Load Balancer, and HTTP clients. 5. CHARACTERIZING SHARED WEB HOSTING WORKLOADS To determine the optimal number of websites which can be supported on Intel Xeon 5500 processor series (dual socket) servers, the Parallels Linux Shared Hosting NG Service Module engineering team investigated web server logs from more than 20 web different servers dedicated to Shared Web Hosting. This data revealed the following key data points: 1% of website sites generate ~60% of the total load on a given web server Requests Distribution by Sites Percent of all requests 100% Requests Distribution by Sites Percent of all requests 100% 50% 50% 0% 1% 2% 3% 4% 5% 6% 7% 8% 9% 10% Percent of sites Percent of sites 1% 2% 3% 4% 5% 6% 7% 8% 9% 10% 85% of HTTP requests are for static content (static HTML pages, picture, etc.) 15% of requests are for dynamic content (mostly PHP scripts) Requests Distribution by Type Requests Distribution by Type 15.00% 15.00% Static content 85.00% 85.00% Dynamic Static content content Dynamic content Web Applications Usage 2010 Parallels. All Rights Reserved. 10 Web Applications Usage

11 Wordpress, joomla and phpbb represent the majority of dynamic requests Web Applications Usage vbulletin 9.04% Other 29.99% Wordpress 23.26% Joomla 13.02% Drupal 1.28% PhotoPost 0.13% Zencart 1.19% XTCommerce 1.04% phpbb 17.91% Gallery 2.79% Egroupware 0.32% Formmail 0.04% Understanding the workload mix of typical Shared Hosting web servers is the first step to determining the number of websites which can be supported per server. The next step is to analyze the number of requests generated per web application type. It was determined that: Average request rate is approximately 50 requests per second for 10K sites and 100 requests per second at peak moments The following diagram details number requests per second over a 6 day period :00 6:00 12:00 18:00 0:00 6:00 12:00 18:00 0:00 6:00 12:00 18:00 0:00 6:00 12:00 18:00 0:00 6:00 12:00 18:00 0:00 6:00 12:00 18:00 0:00 6:00 12:00 18:00 Static (r/s) Other dynamic (r/s) Wordpress (r/s) Joomla (r/s) Based on statistics above, the Parallels Linux Shared Hosting NG Service Module team prepared benchmark system consisting of Set of scripts which use httperf to measure web servers performance A generic average website for benchmarking Shared Web Hosting solutions 2010 Parallels. All Rights Reserved. 11

12 Benchmark scripts take into account typical requests distribution between sites, requests distribution between dynamic/static content, and requests distribution between applications. Two models were selected, which differ only by distribution between web applications used: "Heavy" model "Light" model Wordpress - 30% phpbb - 33% Joomla - 10% vbulletin - 33% phpbb - 60% escene - 33% These graphs represent a model which is very close to the real life scenario found in the web server logs from the site data analyzed from leading web hosting providers. 6. LINUX SHARED HOSTING NG WORKLOAD RESULTS Intel XEON 5570 CLUSTER PURE PERFORMANCE Nehalem-based Test Configuration 7 web-servers (SHNGb1-SHNGb7) 2 x Intel Xeon 5570 CPUs, 16 GB RAM each 100,000 sites on IPs sites per one IP Each site contains several static pages, Wordpress, Joomla and phpbb 1 Load Balancer node (SHNGb10) 4 MySQL servers for site s databases (SHNGb10, SHNGs4, SHNGs5, SHNGs6) 2 x Intel Xeon 5570 CPUs, 16 GB RAM each 4 servers are used as HTTP clients (SHNGs* and SHNGb8) During the test, cluster load was incrementally increased at every 20 seconds measurement period, where the error or time out (4 seconds) was less than 0.1% from all current period requests Parallels. All Rights Reserved. 12

13 Test results In the Heavy Model with the Intel Xeon processor 5500 series servers, the connection response and reply times are key in website density testing; the graph below shows, in transaction requests per second for a 7 server cluster, the point at which the maximum time-to-last-byte becomes unacceptable (4 seconds) is ~2350 transaction requests per second, or ~335 transaction requests per server) min ttflb max ttlb avg ttlb min ttfb max ttfb avg ttfb Intel XEON 5570 and XEON 5470 CLUSTERS - Comparison Xeon 5570-based and Xeon 5470-based Test configurations 3 XEON 5570 web-servers (SHNGb1, SHNGb2, SHNGb3) 2 x Intel Xeon 5570 CPUs, 16 GB RAM each 3 XEON 5470 web-servers (al01, al02, al03) 2 x Intel Xeon 5470 CPUs, 16 GB RAM each 30,000 sites on IPs sites per one IP Each site contains several static pages, Wordpress, Joomla and phpbb 1 Load Balancer node (SHNGb10) 4 MySQL servers for site s databases (SHNGb10, SHNGs4, SHNGs5, SHNGs6) 2 x Intel Xeon 5570 CPUs, 16 GB RAM each 4 servers are used as HTTP clients (SHNGs* and SHNGb8) During the test, cluster load was incrementally increased at every 20 seconds until error or time out requests rate is in range 0.2% to 1% from all current period requests; again, time out is more than 4 seconds for time-to-last byte Parallels. All Rights Reserved. 13

14 Test Results Only the Heavy Model was tested to compare Xeon 5570 servers and Xeon 5470 servers. Based on a maximum time-to-last-byte of 4 seconds, three Xeon 5570 servers can support up to 1,230 requests per second, and three Xeon 5470 servers can support up to 870 requests per second. Requests per second with heavy workload XEON 5570s 3 XEON 5470s 3 XEON 5570s 3 XEON 5470s The Xeon 5570 cluster shows an approximate performance increase of 1.4x with the Heavy Model over the Xeon 5470 cluster running the same workload. 7. BENEFIT TO CLOUD SERVICE PROVIDERS Linux Shared Hosting NG Service Module The Parallels Linux Shared Hosting NG Service Module is a new cloud services offering which addresses the five key areas for hosting services providers that were outlined in the Executive Summary of this white paper. The Linux Shared Hosting NG Service Module solution is delivered as a Parallels Operations Automation Module. The High Density Shared Hosting Service Module can dramatically increase the efficiency of Shared Web Hosting offerings. Cloud service providers can leverage this new capability to increase margins and create a differentiated Shared Web Hosting service. The solution addresses the following limitations found in current offerings to operate a more efficient and profitable Shared Web Hosting service. Lower costs by increasing the density of websites per server The Shared Web Hosting NG Service Module uses Parallels ten years of experience in optimizing computing via Container-based virtualization for the Shared Web Hosting-based servers and uses proprietary modifications to the Linux Apache web server to achieve site densities up to 10,000 sites per server (versus ~1,000 sites for a standard Shared Hosting implementation). To determine the potential savings in opex that could be realized with this increase in the density of web sites per server, the Linux Shared Hosting NG Service Module team developed a Shared Web Services ROI calculator. Using this ROI calculator, one can assess the potential annual savings of Traditional Shared Web Services versus this next generation shared web hosting delivery automation offering. The variables used to calculate the potential ROI includes the following: Current Shared Web Hosting sites under subscription; Annual Churn; Number of Websites per Server; Datacenter Admin Loaded Labor Rate; Number of Person-years Required to Administrate the Web Cluster; Customer Service Representative Loaded Labor Rate; Number of Person-years Required to Deliver 24/7 , Chat, and Phone Support. Based on this, an example ROI calculation could use the following settings: 2010 Parallels. All Rights Reserved. 14

15 The results of assuming a lower churn rate through increased service levels, reduced customer service representative support requests due to a more resilient service offering, and the ability to up sell additional cloud services (such as business class ) would be a net gain of ~$500K over 3 year period when compared to Traditional Web Hosting deployments, as depicted in the graph below: Profit after expenses and fees 2,000, ,800, ,600, ,400, ,200, ,000, , , , , Year 1 Year 2 Year 3 3 year deployment Profit traditional Profit SHNG Using the assumptions above, the ROI calculator also provides a view into infrastructure cost saving obtained through a Linux Shared Hosting NG Service Module deployment, as high website densities have an impact on the number of servers deployed, rack space consumed, and in the number of load balancers and network ports required to support the service. In the example below, server count is reduced from 45 to 6 (or 7.5 to 1). Operating expenses - annual Traditional High Efficiency Infrastructure costs - annual Year 1 Year 2 Year 3 Year 1 Year 2 Year 3 Servers 56, , , , , , Rack cost - 1/24th of 1U rack per server 31, , , , , , Load balancers 2, , , Network ports 2, , , Bandwidth costs 422, , , , , , Infrustructure costs total 557, ,221,43 553, , , , Parallels. All Rights Reserved. 15

16 Increase Average Revenue per User by Offering Additional Services Creating different Shared Web Hosting Quality of Service levels allows Cloud Service Providers to provide guaranteed services and additional revenue streams, as customers are willing to pay for SLA backed Shared Web Hosting services. Parallels Automation also offers additional service modules to provide end-to-end automation for managing domain names, website builder applications, SSL certificates, business class services, Software as a Service, virtual private servers and dedicated servers. Improved service levels lead to a customer experience that reduces churn and retains customers that grow with the Cloud Service Providers. The combination of making it easy for customers to purchase add-ons and up-sells, combined with the ability to easily run promotional campaigns can lead to increased revenue per user. Improve Service Level Management CPU and RAM quotas can be created and enforced on a per site basis to enable different quality of service levels. This capability can be used to force heavy users to pay for heavy resource consumption via a differentiated billing service plan. This ensures such users do not impact the performance of other websites on the server. Cloud Service Providers can also increase service availability and provide differentiated service level agreements while increasing the revenue per user. Parallels leverages specialized software to implement high availability storage clusters to create pools of web servers and storage to minimize the impact of hardware failures. A cloud service provider can create different service levels, each with their own level of redundancy. The premium Shared Web Hosting service level can have mirrored web and storage servers deployed in an active-active web server cluster, while less expensive service levels can be configured with less expensive n+1 redundancy. Increase Predictability of Capacity Planning The Parallels Linux Shared Hosting NG Service Module allows Cloud Service Providers to proactively review their overall Shared Hosting Service and predict when additional resources will be needed to ensure those resources can be purchased in time so as to not impact the end user experience. Automate the user experience and empower the end user Parallels Automation (which consists of Parallels Business Automation, Parallels Operations Automation) automates the user experience to sign up for the service, select their payment method of choice, and provides end-user control panels to manage their service. The cloud service provider also has a completely automated service delivery process, where sites are provisioned in the web farm based on their selected service level and purchased options. Service billing is also automated for the cloud service provider by delivering a consolidated bill per customer for all services. With the Linux Shared Hosting NG Service Module, cloud service provides will be able to offer a differentiated Share Web Service offering (including higher availability and resource guarantees). With a Linux Shared Hosting NG Service Module implementation, Cloud Service Providers can now increase the number of sites per server to reduce costs, and also introduce Quality of Service levels via CPU and RAM quotas. High Availability is also achieved to meet end user expectations while controlling costs. The results of the Parallels Linux Shared Hosting NG Service Module workload testing show that the next generation of Parallels Shared Hosting service offering delivers significant enhancements and performance improvement running on Intel Xeon 5500 processor series-based servers. To learn more about Parallels Linux Shared Hosting NG Service Module, please visit com/products/automation/modules/sharedng. You can also request a consultation with a Parallels Shared Hosting domain expert, who can provide access to the Parallels Automation ROI calculator to evaluate the overall business opportunity to improve the profitability over your current Shared Web Hosting service or to deploy a new offering based on Parallels Automation with the Shared Hosting NG Sevice Module Parallels. All Rights Reserved. 16

17 Intel Recipes for Success Through the Intel Enabled Server Acceleration Alliance (Intel ESAA) program, Intel and Parallels have developed recipes based on Parallels Operations Automation Linux Shared Hosting NG Module. Resellers can get help deploying these solutions using validated configuration recipes from Intel ESAA that combine Parallels software with Intel Server Products. Intel Xeon processor-based server products are engineered from the processor out to meet or exceed global industry standards and enable you to build configurable, customized, and cost-effective solutions. Intel ESAA recipes for Parallels solutions are available for the following Intel Server Products: Intel Server System SR1680MV. A modularized rack server system with two independent dual socket/18dimm nodes in a 1U chassis. Intel Modular Server. A flexible system with integrated storage, computing, and networking. Intel Server Board S5520UR, Intel Server System SR1600UR, Intel Server System SR1625UR, Intel Server System SR2600UR, or Intel Server System SR2625UR. A high-performance, rack-optimized server with a variety of expansion modules for added flexibility. Intel Server Board S5500WB or Intel Server System SR1690WB. Purpose built for energy efficiency and high performance in dense computing environments. Intel Server Board S3420GP, Intel Server System SR1630GP, or Intel Server System SR1630HGP. A cost-effective, 1U rack server featuring the new Intel Xeon processor 3400 series. Learn more: Parallels. All Rights Reserved. 17